IMEKO 22nd TC3, 12th TC5 and 3rd TC22 International Conferences
3 to 5 February, 2014, Cape Town, Republic of South Africa
TORQUE WRENCH CALIBRATION AND UNCERTAINTY OF
MEASUREMENT
J. O.Bangi1, S.M. Maranga2, S.P. Nganga3 and S.M. Mutuli4
1
Kenya Bureau of Standards (KEBS), Nairobi, Kenya, bangij@kebs.org
2
Jomo Kenyatta University of Agriculture Technology (JKUAT), Nairobi, Kenya, smmaranga@yahoo.com
3
Jomo Kenyatta University of Agriculture Technology (JKUAT), Nairobi, Kenya, spnganga@jkuat.ac.ke
4
University of Nairobi, Nairobi, Kenya, stephen.mutuli@uonbi.ac.ke
Abstract: Torque wrenches (TW) are widely used in the
automotive, aviation and building industry. TW are used to
control torque in assembly and fastening operations. Regular
calibration of torque tools guarantees traceability to
international standards. At KEBS, a torque wrench
calibrator with four interchangeable torque transducers and
torque meters are used to calibrate both type I and type II
torque wrenches as per ISO 6789:2003. In compliance to
“Guide for evaluation of uncertainty in measurement”
(GUM), it is a requirement that a complete measurement
result should include its uncertainty, a proposed method
with an example on how to estimate measurement
uncertainties for torque wrenches is presented after taking
into account the various influencing quantities. Uncertainty
values of ± 2.54 N·m, ±1.15 N·m, and ±1.33 N·m are
estimated at 20%, 60% and 100% respectively of a type II
class A torque wrench with a nominal value of 1000 N·m.
The paper aims to provide guideline information for TW
calibration and uncertainty evaluation.
Keywords: Calibration, ISO 6789:2003, Torque Tools,
Torque wrenches, Uncertainty.
1. INTRODUCTION
Torque is one of the important physical parameters, widely
used in several industries such as aviation, automobile,
mechanical, pharmaceutical, power and agriculture.
Applications for torque sensors include determining the
amount of power an engine, motor, turbine, or other rotating
device generates or consumes [1]. Other industrial
applications of torque sensors include the calibration of
torque tools and sensors; measuring peel forces, friction, and
bottle cap torque; and testing springs. The importance of the
torque measurements is to control the fastening force on
assembly parts and to measure mechanical power. Torque is,
generally, measured by mechanical type torque wrenches/
meters. A torque wrench is a tool used to precisely apply a
specific torque to a fastener such as a nut or bolt. Torque
screwdrivers and torque wrenches have similar purposes and
mechanisms [2]. Regular calibration of torque tools
guarantees traceability to international standards.
Calibration procedures for hand torque tools are defined in
the ISO 6789:2003 standard [3]. ISO 6789:2003 however,
does not provide any information or guidelines on how to
evaluate the uncertainty of measurement [4]. On the other
hand, according to GUM [6], it is necessary to report a
complete measurement result including its uncertainty. Also
IEC/ISO 17025:2005 standard [8] requires that a calibration
laboratory, performing its own calibrations, shall have and
shall apply a procedure to estimate the uncertainty of
measurement. This paper presents the classification and
calibration, and suggests a measurement uncertainty budget
for torque wrenches as per GUM. The aim of this paper is to
provide guideline information for TW calibration and
uncertainty evaluation to quality managers with a view to
encouraging setting up a torque wrench calibration facility
in their own organisations.
2. CLASSIFICATIONS OF HAND TORQUE TOOLS
According to ISO 6789:2003, hand torque tools are
classified as follows: Type I – indicating torque tools and
Type II – setting torque tools. Below is a description of
Type I and Type II.
2.1. Type I: Indicating torque tools
These are tools that indicate by means of a mechanical scale,
dial or electronic display, the value of torque exerted by the
tool at the output drive [3]. The classes in type I are;
• Class A: wrench, torsion or flexion bar
• Class B: wrench, rigid housing, with scale or dial
or display
• Class C: wrench, rigid housing and electronic
measurement
• Class D: screwdriver, with scale or dial or display
• Class E: screwdriver, with electronic measurement
Below is a description of a few of type I wrenches;
Dial indicating torque wrenches: You just turn until
the dial indicates the torque you want (See Fig. 1).
 on the internal spring can be responsible for a long-term
decrease in accuracy. It is necessary, before performing
calibration, to keep the torque at the minimum setting for a
certain amount of time (12 hours), in order to exclude or
reduce this effect [9]. Even so, it should be calibrated
regularly if used for critical parts like suspension and
internal engine fasteners.

Fig1: Dial indicating torque wrench [5]

Bending beam torque wrenches: The large centre

beam bends as you apply torque, while the unbending
pointer beam lets you read the torque directly. When the
desired torque is reached (see Fig. 2), the operator stops
applying force. The disadvantages are that your eyeball has
to be parked directly above the pointer while you read the
scale, which is tough in hard-to-reach places and calibration
cannot be performed automatically, since the operator has to
look at the indicator and write manually the value. Fig 3: Click type torque wrench [2]

Fig. 4: shows the inside structure of a click TW, with a

multi-lever action. The spring in the hilt applies a force
against one or more elements, which are connected, through
a multi-lever system, to the square drive. Setting the TW
means compressing the spring and applying a higher force to
the elements. When the user operates the TW and the force
balance is reached, there is a click point (peak value) and an
immediate decrease of the torque value [9].

Fig. 2: Beam type torque wrench [2].

2.2. Type II: Setting torque tools

These tools are pre-set to a certain torque value, so that

when the prescribed value of torque is exerted by the tool at
the output drive, a signal is released (e.g. audibly, visibly,
and perceptibly) [3]. The classifications in type II are;
2.2.1 Adjustable graduated torque tool (Type II, Class
A, Class D and Class G)
These tools are designed to be adjusted by the user, which
have a scale or a display to assist adjustment.
2.2.2 Adjustable non-graduated torque tool (Type II,
Class C and Class F)
These tools are designed to be adjusted by the user with the
aid of a calibration device
2.2.3 Torque tool with fixed adjustment (Type II, Class
B and Class E)
Tools not designed to be adjusted by the user, i.e. having a
single setting. Below is a description of a few of type II
wrenches;
Fig. 4: Multi-lever inside mechanism of a click torque wrench [9].
Slipper torque wrenches: A variant of the previous
one, this TW type cannot be adjusted directly by the user, as
it doesn’t have a setting scale or indicator. It is also known
as “production type TW”, and is normally set against a
torque wrench analyzer or Torque Wrench Calibration
Device (TWCD) (see Fig. 5) and then dedicated to a
particular application, where modifications of the value of
the nominal torque are not frequently required. From the
metrological point of view, this is the “safest” tightening
system, as it eliminates over-tightening and discourages
unauthorized alteration. It has a smooth decreasing curve
after the peak value [9].

Click torque wrenches (adjustable by the user): This

pro-grade tool is preset to the desired torque and will click
tactilely and audibly when it reaches the desired torque (See
Fig. 3). At the point where the desired torque is reached, the
clutch slips, signalling the desired torque and preventing
additional tightening. It should always be turned down and
stored at the lowest possible setting. The continuous “stress” Fig. 5: Slipper torque wrench (production type) [2]
 Electronic torque wrenches: For electronic torque
wrenches (see Fig. 6) measurement is by means of a strain
gauge attached to the torsion rod. The signal generated is
converted by the transducer to the required unit of torque
(N·m, lbf·ft etc.) and shown on the digital display. A
number of different joints (measurement details or limit
values) can be stored. These programmed limit values are
then permanently displayed during the tightening process by
means of LEDs or the display. At the same time, this
Fig. 8: Automatic Torque Wrench Calibrator used at KEBS.
generation of torque wrenches can store all the
measurements made in an internal readings memory. This 3.1 Evaluation of the Calibration result
readings memory can then be easily transferred to a PC via
The evaluation of the deviation shall be made by the
the interface (RS232) or printed straight to a printer. [2]. following formula [3]:

Fig. 6: Electronic torque wrenches [2]
3.0 CALIBRATION OF TORQUE WRENCHES
Where:
As (%) is the calculated relative deviation of the torque tool;
is the indicated value of the torque tool;
is the reference value (determined by the calibration
device).

Calibration shall be carried out at a temperature fluctuating In the special case of Type II classes C and F tools (without
by not more than ± 1 °C. This temperature shall be in the scale), the value of is calculated as mean of the ten
range between 18 °C to 28 °C (maximum relative humidity reference torques measured by the calibration facility.
90 %) and shall be documented [3]. In general, the This value is then used in formula (1) [4].
calibration procedure consists of one and five pre-loadings
for Type I and Type II respectively with the maximum 3.2 Tolerances
torque without recording the values, followed by testing first
at 20 % and then at about 60 % and at 100 % of the According to ISO 6789:2003, permissible deviation of the
maximum torque value of the torque tools (or at the torque value indicated by the tool, from the simultaneous
nominal/set value for tools of Type II, Classes B and E). All indication of the calibration device, is as given in Table 1-3
readings shall be within the maximum tolerances specified below [3];
in ISO 6789:2003 and shall be recorded. The conformity can
be declared if the deviations are within the limit set by the Indicating torque tools (Type I)
standard (± 4%, respectively ± 6% for some classes). There
is no measurement uncertainty taken into account in Table 1: Permissible deviation (Type I)
contradiction to the IEC/ISO 17025:2005 standard which
Classa Maximum torque Value
requires that conformity statements must be made including
≤ 10 N·m > 10 N·m
the uncertainty of measurement [7]. This is very important A and D ±6%
for accredited calibration laboratories [4]. Calibration can be B, C and E ±6% ±4%
performed with automatic or manual calibration devices.
See Fig. 7 and Fig. 8 for manual and automatic TWCD. Setting torque tools (Type II)

Table 2: Permissible deviation (Type II, Classes A, B, D, E and G)

Classa Maximum torque Value

≤10 N·m >10 N·m
A and B ±6% ±4%
D, E and G ±6% ±6%

Table 3: Permissible deviation (Type II, Classes C and F)

Class Maximum torque Value

≤10 N·m >10 N·m
C ±6%± ±4%
F ±6%

For Classes C and F, the torque value set is equal to the arithmetical mean
Fig.7: Manual calibration device (digital meter) used at KEBS
of the 10 test readings
 4.0 EVALUATION OF THE UNCERTAINTY OF necessary if that particular calibration position is not
MEASUREMENT included in the certificate.

Uncertainty of measurement shall be calculated in Table 4: Results from a Calibration Certificate [7]
accordance with the GUM [6]. The GUM Steps are;
Clockwise
Actual Average Output Relative Transducer
 Model the measurement
Applied transducer using expanded Class
 Identify and quantify the sources of uncertainty Torque readings fitting uncertainty
 Categorize as type A or type B equation (k=2)
 Manipulate appropriately to obtain N·m mV/V mV/V ± %.
0 0.0000 0.0000 0.0000
i. Standard uncertainties, u(xi) 200 0.1004 0.1003 0.519 1.0
ii. Sensitivity coefficients, ci 400 0.2005 0.2006 0.353 0.5
iii. Uncertainty contributor, u(yi) 600 0.3008 0.3009 0.241 0.5
800 0.4012 0.4011 0.416 1.0
 Combine to obtain combined standard uncertainty, uc(y) 1000 0.5017 0.5013 0.254 0.5
 Expand to obtain an expanded uncertainty, U 1200 0.6012 0.6015 0.365 1.0
 Report the result 1400 0.7018 0.7017 0.283 1.0
1600 0.8019 0.8018 0.238 0.5
4.1 Measurement Model 1800 0.9018 0.9020 0.306 1.0
2000 1.1021 1.1020 0.258 0.5
The calibration result is the output signal of the torque
measuring device and is obtained from the approximate Contribution of the resolution (ur): This is due to the
mathematical model [10]: resolution of the TW scale divided by a standard value. Each
deflection value is calculated from two readings (the reading
(5) with an applied torque minus the reading at zero torque).
Because of this, the resolution of the indicator needs to be
Where: included twice. This is the range of the rectangular
Influence of resolution of the TW, r distribution. Consider a TW with a resolution of 2 N·m and
Influence of repeatability error, b’ therefore the semi-range is (2 N·m /2) =1 N·m. The divisor
Influence of the relative uncertainty for rectangular distributed uncertainty contributors is √3 [6].
of the values of the reference torque transducer.
Contribution of the TW repeatability (u ): This result
4.2 Identifying the sources of uncertainty from the variability in the measurement results obtained
after repeating the measurement five (or ten) times. It can
be dealt with either as “repeatability” or as
“reproducibility”. This contribution is considered as a type
A uncertainty and is not divided by the coverage factor.

Environmental factors: Too extreme temperatures and

humidity should be avoided [4]. Temperature variation less
than ± 1 °C and humidity lower than 90% during calibration
Fig. 9: Sources of uncertainty for TW. should be maintained [3].

4.3 Quantifying the sources of uncertainty 4.4 Categorize as type A or type B

The following contributions can generally be taken into • - type B, not statistically determined
account: • - type B, not statistically determined
• - type A, statistically determined (standard deviation)
Contribution of the Reference torque (utcd): This is the
uncertainty due to the torque transducer of the TWCD. This 4.5 Manipulate appropriately to obtain
value should be taken from the manufacturer’s calibration
certificate, for the corresponding measuring range or defined i. Standard uncertainties, u(xi)
position, divided by the coverage factor (usually k=2). This ii. Sensitivity coefficients, ci
uncertainty component is a combination of influences from iii. Uncertainty contributor, u(yi)
repeatability in unchanged mounting position,
reproducibility in different mounting positions, deviation • Combine to obtain combined standard uncertainty, uc(y)
resulting from the fitting curve and instrument resolution. • Expand to obtain an expanded uncertainty, U, at an
As an example (see Table 4) 0.52% is the actual value appropriate level of confidence
corresponding to the 200 N·m position. Interpolation can be
 4.6 Uncertainty budget
The standard uncertainty expressed in units of
indication and the relative standard uncertainty is The expanded uncertainty of measurement U or the
obtained by the law of propagation of uncertainty in the expanded relative uncertainty of measurement W for each
approximation of non-correlated variables [11]: calibration step is calculated from the combined uncertainty
of measurement, using equation (10 or 11), with coverage
factor k= 2:

Table 5 below contain the symbols and designations used in

this paper.
With

Table 5: Symbols, designations and units

Symbol Designation Unit

applied calibration torque N·m
indicated torque values N·m
mean value of the torque measuring device of increasing steps N·m
’ repeatability N·m
b reproducibility N·m
n number of increasing series
j index of selected series N·m
r resolution of the torque tools N·m
uncertainty contribution of resolution N·m
relative uncertainty contribution of resolution %
’ uncertainty contribution of repeatability N·m
’ relative uncertainty contribution of repeatability %
uncertainty contribution of the applied torque value of the torque standard devices N·m
relative uncertainty contribution of the applied torque value of the torque standard devices %
combined standard uncertainty of measurement N·m
combined relative standard uncertainty of measurement %
U expanded uncertainty of measurement N·m
W expanded relative uncertainty of measurement %
standard uncertainty of measurement N·m
relative standard uncertainty of measurement %

Table 6: Proposed distributions and corresponding standard uncertainties for important parameters

Quantity Evaluation of Distribution Standard uncertainty u relative standard

standard uncertainty function in N· m uncertainty w , %
Repeatability, b’ type A Rectangular
’

Resolution of the Type B Rectangular

torque tools r

Reference torque Type B Normal

Combined standard uncertainty ’

Combined relative standard uncertainty

Expanded uncertainty ,k=2
Expanded relative uncertainty of measurement ,,k=2
 Table 7: Sample calibration data of setting torque wrench type II class A with nominal value of 1000 N·m,
resolution: 2 N·m

Set Readings in N·m

Readings readings Average Std dev of
(%) in N·m 1 2 3 4 5 readings readings (%)
20 200 201.5 201.2 199.8 201.2 199.8 200.7 0.82
60 600 599.7 599.3 599.4 599.3 600.1 599.5 0.35
100 1000 1002.8 1003.1 1002.0 1003.2 1002.0 1003.0 0.45

Table 8: Example of uncertainty budget of setting torque wrench type II class A at 200 N·m, resolution 2 N·m

Quantity Estimate Relative Standard Probability Sensitivity Relative uncertainty

uncertainty W (Xi) , Distribution coefficient Ci contribution Wi (y) , %
Xi, N·m %
200
200.7
0.58 Rectangular 1 0.58
’ 0.82 Normal 1 0.82
0.52 Normal 1 0.52
Combined relative uncertainty, 1.27
Expanded rel. uncertainty, W at k =2 2.54

From the sample calibration data in Table 7, uncertainties are calculated according to Table 6 and are shown in Table 9.

Table 9: Calculated uncertainties

in N· m
Readings, % Set readings ’ U
20.00 200.00 0.52 0.82 0.58 1.27 2.54
60.00 600.00 0.24 0.35 0.58 0.58 1.15
100.00 1000.00 0.25 0.45 0.58 0.66 1.33

4.6 Report the result
The uncertainty of measurement of the torque wrench at the
measuring position of 200 Nm would then be given as 200.7
± 2.54 Nm, at a Level of Confidence of 95,45%.
5.0 CALIBRATION CERTIFICATE AND DURATION
OF VALIDITY
Where the calibration of a torque measuring device has
satisfied the requirements of the standard, the calibration
laboratory should draw up a certificate stating the following
information [11], in addition to that shown in guide EA-4/01
[12].
a) Identify all the elements of the torque measuring
device and its components, including mechanical
coupling components to the calibration equipment;
b) The method used, identifying whether clockwise
and/or anti-clockwise torque, incremental and
together with reference to ISO 6789:2003 standard;
c) The resolution of the torque measuring device;
d) The temperature at which the calibration was
performed;
e) A statement on the expanded uncertainty of
measurement and the equation of the
fitting curve where applicable;
f) Where required, a statement regarding the
conformity of the calibration results to a
particular classification at the criteria used
5.1 Re-calibration of torque wrenches
The calibration interval shall be chosen on the basis of the
factors of operation such as required accuracy, frequency of
use, typical load during operation as well as ambient
conditions during operation and storage conditions. The
interval shall be adapted according to the procedures
specified for the control of test devices and by evaluating
the experience gained during recalibration. If the user does
not utilize a control procedure, a period of use of 12 months,
or approximately 5000 cycles, can be taken as a default
value for the recalibration interval. This period should be
reduced when the tool is frequently used. The TW shall be
recalibrated when it has been subjected to an overload
higher than that applied in the overloading test, after repair
or after inexpert handling, which may have an effect on the
uncertainty of measurement [3].
6. CONCLUSIONS
This paper presents the classification and calibration of
torque wrenches as per ISO 6789:2003 standard. In
compliance to GUM, a method of determining important
and relevant influencing quantities in torque wrench
calibration is described. Example of uncertainty budget of
setting torque wrench type II class A is presented and the
expanded relative uncertainty values were found to be ±
2.54 N·m, ±1.15 N·m, and ±1.33 N·m at 20%, 60% and
100% respectively of the nominal value of 1000 N·m.
5. REFERENCES
[1] Source:
http://www.omega.com/literature/transactions/volume
3/ force3.html.
[2] Source: http://en.wikipedia.org/wiki/Torque_wrench
[3] ISO 6789:2003, Assembly tools for screws and nuts –
Hand torque tools – Requirements and test methods for
design conformance testing, quality conformance testing
and recalibration procedure.
[4] Röske. D.; “ISO 6789:2003 Calibration Results of
Hand Torque Tools with Measurement Uncertainty – Some
Proposals”, XX IMEKO World Congress, Busan, Republic
of Korea September 9-14, 2012. CD-ROM, file name:
594_F_N_TC3_279_297.pdf.
[5] http://www.belknaptools.com/dial-torque-wrenches/dial-
indicating-torque-wrenches/
[6] Evaluation of measurement data - Guide to the expression
of uncertainty in measurement, Internet:
http://www.bipm.org/utils/common/documents/jcgm/JCG
M_100_2008_E.pdf, accessed: 2013-10-15.
[7] National Institute of Standards (NIS) Egypt, Certificate
of calibration No. 562/14/2013, 2013.
[8] ISO/IEC 17025:2005, General requirements for the
competence of testing and calibration laboratories.
[9] L. G. Bochese, “Uncertainty of Measurement in Torque
Wrench Calibration”, NCSL International Workshop and
Symposium, Tampa, Florida, USA, August 17-21, 2003.
[10] Saenkhum N., Torque wrench calibration devices
according to DKD-R 3-8: 2003, 2011
[11] EURAMET cg-14, Version 2.0 - Guidelines on the
Calibration of Static Torque Measuring Devices, 2011.
[12] EA-4/01: Requirements Concerning Certificates Issued
by Accredited Calibration Laboratories, edition 1,
November 1995.